The thoracic artery in P. subis arose either from the subclavian artery, or from the coracoid artery. Likewise in other swallows, both of these origins were found. In doves the thoracic artery arose consistently from the pectoral stem, lateral to the origin of the axillary artery.
The coracoid artery in P. subis and other swallows arose from the subclavian artery, either opposite the base of the axillary artery, or medial to that vessel. In all doves studied the coracoid artery arose from the subclavian medial to the axillary artery. I observed much individual variation in the branches of the coracoid artery (that is to say, in the vessels of the coracoid complex). In S. inca the sternal artery arose either from the coracoid artery, or independently from the subclavian. As mentioned earlier, in members of both families the vessels to Mm. coracobrachialis posterior and subcoracoideus are highly variable, arising in swallows from the coracoid artery or from the subclavian artery, and in doves from either of these two sites or from the axillary artery. The distribution of these arteries after their origin is also diverse.
Individual variation in the arteries of the thorax has been recorded previously. Bhaduri, Biswas, and Das (1957:2) state that, in the domestic pigeon, "the origin and course of various smaller arteries... show noticeable variation," although they do not specifically state to which vessels they are referring. Fisher (1955:287-288) found variability in the Whooping Crane, Grus americana, of the axillary, coracoid, thoracic, and pectoral arteries. In one specimen he found these vessels arising on the right side from the subclavian, in the sequence just listed, and on the left side all arose from the same point. Berger (1956:439-440) strongly emphasized the variability of the vascular system, calling it the most variable in the body. As he stated, this high degree of individual variation seems to be due to the embryological development of the system, wherein many of the adult channels of circulation are derived from embryonic plexuses.
Intrafamilial Differences
In spite of the rather extensive amount of individual variability in some vessels, I found the over-all pattern of arteries to be relatively constant within the family Columbidae and within the family Hirundinidae. There are, nevertheless, several intrafamilial differences needing some further discussion and clarification.
Others have reported the occasional presence of more than one coracoid artery on each side in some columbids, these arteries being described as arising from various sites and being variously named. Bhaduri and Biswas (1954) described the arterial situation in seven species of the family Columbidae (Columba livia, Streptopelia tranquebarica, S. chinensis, S. senegalensis, Chalcophaps indica, Treron bicincta, and T. phoenicoptera) and stated (op. cit.: 348) that "The sternoclavicular [= coracoid] artery is similar in all the species, but the domestic pigeon seems to be unique in that it has, in addition, a small vessel, the accessory sternoclavicular." This artery was described later, in the domestic pigeon, as follows (Bhaduri, Biswas, and Das, 1957:5): "A minute and insignificant vessel which has been termed the accessory sternoclavicular artery... is given off close to the origin of the sternoclavicular. It passes anteroventrally to supply the adjacent muscles." Glenny (1955:577) described the arterial pattern characteristic of members of the family Columbidae (more than 30 species studied by him) and stated that "three pairs of coracoid arteries are found in Otidiphaps nobilis, normally one or two pairs may be found." As suggested by Bhaduri and Biswas (1954:348), the "accessory" vessel probably corresponds to a vessel previously described by Glenny (1940) in Streptopelia chinensis and referred to as the "coracoid minor."
Bhaduri and Biswas (1954:348) have suggested that "the accessory sternoclavicular artery occurring sporadically as it does in some species of diverse groups may not have any phylogenetic value."
In no case did I find more than one coracoid artery on a side. When one of the highly variable arteries feeding Mm. coracobrachialis posterior and sternocoracoideus (arteries 4 and 4a, Fig. 7) arises from the subclavian or axillary artery instead of from the coracoid artery, that vessel may have been interpreted by others as a second (accessory or minor) coracoid artery. If so, this artery probably does not "occur sporadically." Rather, its origin from the subclavian, axillary, or thoracic artery may be sporadic, subject to individual variation, and it may have been overlooked when it arose from the coracoid artery.
Of the vessels described here, the only one that differed distinctly in one species was the sternal artery. In Scardafella inca the right sternal vessel was long, extending down the mid-line of the inner surface of the sternum, whereas in other columbids the right and left arteries ramified on the anterior part of the inner surface of the sternum, or were altogether lacking. I am unable to account for the differential development of this artery in S. inca.
In describing the arterial arrangement in the seven species of Indian columbids named earlier, Bhaduri and Biswas (1954:348) state that all species except Treron phoenicoptera have two "internal mammary" arteries on each side "showing variable sites of origin." These arteries were later described (Bhaduri, Biswas, and Das, 1957:4-5) as "a slender (outer) internal mammary artery... to the outer wall of the thoracic cavity" and "a slender (inner) internal mammary artery... to supply the inner wall of the chest cavity." From this description, the question arises as to whether the "outer" one of these arteries should properly be called an external instead of internal mammary artery. In any case, I saw no specimen possessing two thoracic arteries on a side.